Two million preliminary diagnoses made by Moscow therapists using a neural network

August 17
Social sector

The capital therapists have made two million preliminary diagnoses during doctor’s visits through a service that uses a neural network.

"Today, the routine practice of doctors in all adult polyclinics features the use of modern digital tools such as, for example, a medical decision support system, which includes making a preliminary diagnosis through the use of artificial intelligence. The digital assistant has already helped therapists make two million preliminary diagnoses. The service provides for higher accuracy of diagnostics, since the patient's complaints are addressed by both the doctor and AI. The tool also enables a doctor to unify the filling out of the visit record on most common diseases," the press service of the City's Social Development Complex informed.

The doctor’s digital assistant carries out analysis using a machine learning algorithm created at AI laboratory. It analyzes the patient's complaints that the doctor has entered into the system and then offers the therapist to choose one of the three most likely diagnoses, while not limiting the doctor in making a final decision, meaning that the therapist can choose one of the proposed diagnoses or make an independent one.

Higher accuracy and time savings

The neural network is improving its capabilities in the diagnosis of diseases. The service has already become more accurate by 7.4 percent, the model accuracy amounting now to 68 percent. After making a preliminary diagnosis, the doctor sends the patient to diagnostics and appoints relevant tests or consultations with subspecialists. At this stage, the medical decision support system suggests using another service – packaged appointments with the required set of laboratory and instrumental examinations, as well as consultations with subspecialists to confirm the preliminary diagnosis. The doctor may, at his/her own discretion, appoint additional diagnostic procedures or reduce the list of those suggested if the patient has recently undergone them. The doctor can also see this information in the Unified Medical Information and Analytical System (UMIAS).

The use of the medical decision support system provides for a 10-time reduction of the average time required for diagnostic tests appointment. The development of digital services reduces the doctors’ routine work and draws their attention to the important details in the patient history that might be missed if not analyzed by the neural network. The service also prevents repeated appointments of the same examinations.

The medical decision support system is based on the clinical protocols developed by the main specialists of Moscow jointly with federal experts on the basis of clinical recommendations and international standards.

Implementation of such solutions became possible thanks to the unified digital healthcare platform that was created and has been developed jointly by the Moscow Social Development Complex and Moscow Department of Information Technologies. AI-aided services are also integrated into the unified radiological information service of EMIAS and are used in radiology. They analyze radiology results and expedite the search for the signs of diseases on medical images by doctors. New technologies are also used for diagnosing a number of lung pathologies, breast cancer, detecting COVID-19 and other diseases.


If you continue to use our website, you are agreeing to accept the use of cookies on your device. Cookie files ensure the website’s efficiency and help us provide you with the most interesting and relevant information. Read more about cookie files.
Accept ccokies